Someone posted Feynman’s personal appendix to the Rogers Commission Report (the report on the failure of Challenger) to Hacker News this morning. I’ve read it several times, but this line stood out at me this morning:

The proper way to save money is to curtail the number of requested changes, not the quality of testing for each.

It was found that while the hardware engineering practices were directly to blame for the loss of the vehicle and her crew, the software engineering practices were top notch. There was growing pressure to reduce the cost of changes to the space shuttle’s computer software, and this was Feynman’s response to that attitude.

It strikes me as wise, even if most of us aren’t sending people to space. We should constantly be aware of what it costs to change the plan and not sacrifice our principles to do on-budget what we, or even our client’s, want.